accuracy of an ADC

log2(Error) =3D Bits

table:

Error^-1, n'est ce pas?

yes or a minus on the log2()

Thanks for the help... where can i find an explanation on this formula? I haven't foud this formula before...

I will explain it.

Normally you think of binary numbers as having bits to the left of the "decimal point". The bit values you give them are 1, 2, 4, 8, 16, 32 and so on. For thinking about ADCs it is sometimes better to take a different view.

Lets call the full scale of the ADC 1.0V so that we don't have to have a constant for the full scale.

Now consider the simplest ADC, which is a one bit ADC. The one bit indicates that Vin is greater than 0.5V

Now consider the two bit ADC. The most significant bit is still

As we add more bits, the least significant goes 1/2, 1/4, 1/8, ...etc.

We can say that the LSB is always one over 2 to the power of N.

Y =3D 2^N

Now do a little math to get N alone.

N =3D log2(Y)